- Tektronix, winner of the DesignCon 2023 Best Paper Award, is recognized for its contribution to the diverse educational objectives of DesignCon, the American event dedicated to the design of electronic chips, boards and systems.
The paper was written by Tektronix team members Wenzheng (Shawn) Sun and Muhammad Saad Chughtai in collaboration with Yongjin Choi and Chris Cheng of Hewlett Packard Enterprise (HPE). Priyank Kashyap and Pual Franzon of North Carolina State University also co-wrote the paper. The Best Paper Awards serve to acknowledge the authors who receive them as leading practitioners in semiconductor and electronic design, and this submission was paired with a corresponding proof of concept demonstration at the Tektronix DesignCon booth.
The demonstration setup consisted of a 50Gbit/s PAM4 source, a variable ISI channel and a Tektronix Série DPO70000 SX oscilloscope, which was then used to showcase a live percentage loss injection, acquisition, software EQ and Eye diagram accumulation using PAMJET. The PAMJET processing took less than one minute, and the machine learning model showed instant eye diagram generation a feat that is many times faster, less intensive, and more repeatable than current techniques. This breakthrough uses innovative technology coupled with bespoke machine learning algorithms and has many potential applications including Data Center improvements, Transceiver anomaly detection, and further breakthroughs within Artificial Intelligence.
“It’s an honor to receive recognition for our cutting-edge technology at Tektronix, but this award is especially important as it recognizes Tek’s contribution in the Test and Measurement sector by showing instant eye diagram generation,” shares Wenzheng (Shawn) Sun, Tektronix Software Design Engineer. “Our paper, Data-Driven PAM4 SerDes Modeling & Generative Adversarial Network, showcases a significant step forward in Tek’s innovation, as both eye diagrams were visibly identical in our demonstration. This research was done in partnership with HPE and North Carolina State University.”
“HPE is proud to have worked with Tektronix on this significant project,” shares Chris Cheng, Distinguished Technologist at Hewlett Packard Enterprise. “Having the chance to deliver this paper and corresponding demonstration showing instant eye diagram generation on previously unseen loss conditions and EQ settings through a cutting-edge machine learning model is both exciting and significant for our industry.”
Recipients of DesignCon’s Paper Award are selected through a two-prong process. The first step is a review of the full-length papers accepted by the members of the DesignCon Technical Program Committee, who then rank and determine the finalists based on quality, relevance, impact, originality, and lack of commercial content. Winners are then chosen from the finalists based on the quality of their presentations at the conference as rated by the attendees.